"Guess my dog's breed"... Maybe?

The popularity of breed identification tests has been increasing in parallel with the interest in human “ancestry” genetic tests. The technology is similar, as are the sometimes surprising results. We recently had contact from a friend of IPFD who put their dog Olive through three different breed identification tests – with some expected, and unexpected (and perhaps concerning) results. This blog explores some of the reasons for variation in results, as well as any benefits and risks to using these tests. 

What is the purpose of these tests and what did you hope to gain?

For many people, breed estimation testing is mostly a fun experience, trying to put some named breeds behind the traits in a mixed or unknown breed dog. Different test providers will claim to provide information including estimated size, likely behaviors, physical traits, “relatedness”/family trees, and disease risks. No test provider currently indicates pedigree or breed “purity” as it relates to registration with a kennel or breed club. 

Before undertaking any genetic testing, it is important to take some time to identify the purpose of testing, and what information you are wanting. There will be different priorities for different users, and not considering these early on will make selecting the right product(s) difficult. Those new to testing might want to review an article on what genetic test types are available. 

Olive's Experience

Olive is a medium-sized dog from Canada, who is of unknown origins. Her owners decided to test Olive using 1 breed identification test from Canada, and 2 tests from the USA. The tests all claimed to estimate the likely breeds composing Olive, and to varying degrees provide some information about those breeds, and possible traits for Olive. The results of all three tests were provided to IPFD by the owner, as they were surprised by the differences in results. The two USA test results were fairly similar. Both use databases that are optimized for mixed-breed dogs, and not purebred/pedigree dogs, which by definition are more inbred and could skew the information. This is important to bear in mind if you have a purebred dog and are attempting to confirm its breed – with few exceptions, these tools are not designed to do that! The two USA test providers offered similar number of relevant markers, but each have different reference panels of dogs for those markers. Generally, the more markers the better, but it is also important that the selected markers are breed-relevant, and in many cases, population (e.g. country/continent) relevant. For Olive, the results estimated in the first few generations – e.g. her “closest” relatives, from both test providers were either the same breed/type or similar, mainly shepherds and herding breeds. There were, however, some wild suggestions (and fewer in common between the test providers) for dogs “further back” in Olive's family tree. This is likely due to the differences in markers being screened for, and the different reference panels used, and how they assign markers in defining specific breeds. For example, one provider set a threshold in their panel as Olive being German Shepherd dog and accounted for minor shepherd-like variations within a general mixed-breed term, whereas the other provider chose to specify Dutch, German, and White Swiss dog mix. Both are likely accounting for the same thing, genetically speaking. Practically speaking they are indicating similar degrees of precision about similar breed-related traits. 

The wildcard was the Canadian test. These test results seemed to represent breeds they believed to be in very near generations – roughly parents (50%), grandparents (25%), and great-grandparents (12.5%). This, in principle, isn’t a bad approach. All test providers had information indicating that the breed-related information in near generations was likely to be the most relevant to Olive. What was curious about these results, however, was that they were not in keeping at all with the other test providers nor in correlation with any distinctive breed-related traits observed in Olive. The reasons are a bit of a mystery, but likely due to this provider using a smaller number of markers, and variations in their breed reference panel. While in this case unlikely to impact Olive, it is something to think about if your aim was to identify breed-associated risks or traits. 

How much could precision matter?

If your primary interest is the novelty element of breed estimations, precision may be less of a concern. However, many people utilizing these tests are also looking for deeper information. In this case, choosing the test provider with the correct genetic tool for your priorities and level of precision is critical. If a test provider is using a relatively small number of markers, and if they are poorly associated with a wide variety of well-defined breeds, then the differences between test providers could be like comparing a microscope to a telescope. More markers aren't a guarantee of precision without a good reference panel, but there needs to be enough markers to have clear breed associations to make distinctions between breed and breed types. 

Potential benefits to breed estimation:

• identification of potential breed-associated health risks (varies by product)
• early awareness of breed-associated risk factors may help owners choose preventative lifestyle changes, or catch health concerns earlier
• puppies or young dog traits such as size estimations might help in homing appropriately
• breed-associated behavioral traits may be useful for dogs with unknown origins
• "relatives" finders or related dogs on some products can be fun and contribute to human-animal relationship

These benefits could be helpful in preventative care, or in supplementing veterinary investigations. Dog shelters are keen to be able to provide whatever information they can on dogs with unknown breed origins – particularly size and behavior – which are derived from the breed typing, rather than specific associated genes. 

Considerations and risks:

• with poor precision there is a risk of misdiagnosis, or inappropriate veterinary treatment for disease risks
• breed identification can bring challenges associated with the restriction or banning of certain breeds. Having a banned breed or breed-cross with “too much” wolf, bully, mastiff, etc. can impact the ability to get an apartment, find homes for shelter dogs, or in extreme cases, ownership may be illegal according to local policies
• test accuracy and information can vary dramatically between providers
• just like any genetic testing, breed testing isn't a guarantee of health or a health certificate
• these products do not determine if a dog is purebred or pedigree (nor do they claim to!)

In Summary

If people are using any of these tests as a fun novelty of what breeds their dog is composed of, to play spot-the-trait in the suggested breeds, then the provider may not matter too much - as in human genetic ancestry tests, where everyone thinks they are Neanderthal or Cherokee. 

However, if you are wanting to know about actual health risks or traits that could impact welfare, then it is worth doing some research to find a provider that is able to have the precision in breed/type estimation, and the tests for traits that are important to you and your dog. In most cases, breed-associated traits beyond 2-3 generations are less likely to be precise enough to be informative, particularly if you have a dog that is truly a mix. 

In Olive's case, one health/welfare consideration where the breed testing may have helped is the identification of a number of shepherd and collie breeds that are likely part of her heritage. In her case, MDR-1 genetic testing might be prudent as there are multiple shepherd/collie breeds where this is a relevant test, and it is a risk that is important for the dog's welfare. Olive's veterinarian would also likely guess shepherd/collie breeds, and act accordingly... but what if she didn't have such distinctive, visible traits?

As for her breeds? Our guess is a shepherd-collie-very-good-girl. 
 

For more information:

Breed identification tests

Searchable breed-specific genetic testing, and information on genetic test providers can be found via HGTD.

Title image with thanks to Brett Jordan via Pexels. Photos of Olive were kindly provided by her owners.